compiler bug in 1.9.9

compiler bug in 1.9.9

[Daniel Llorens del Río]

This is with the latest from git.

(use-modules (srfi srfi-26))
(define (f! A B) (array-fill! B 0.))
(define x (make-typed-array 'f64 0. 2))
(cut f! #2f64((0 1) (2 3)) <>)

Throw to key `wrong-type-arg':
ERROR: Wrong type (expecting array): #<<variable-cache-cell> key: f!>
Entering the debugger. Type `bt' for a backtrace or `c' to continue.
0 debug> bt
In ../../src/module/ice-9/boot-9.scm:
  1042: 12 (#<procedure e04f90 ()>)
In ../../src/module/system/repl/common.scm:
   145: 11 (repl-eval #<<repl> language: #<<language> name: sche...> #)
In ../../src/module/system/base/compile.scm:
   237: 10 (compile (cut f! #2f64((0.0 1.0) (2.0 3.0)) <>) #:from # ...)
   183:  9 (lp (#<procedure compile-asm (x e opts)> #<proce...> #)  
# ...)
In ../../src/module/language/glil/spec.scm:
    31:  8 (compile-asm #<glil (program () (source  
((breakpoint ...> ...)
In ../../src/module/language/glil/compile-assembly.scm:
   155:  7 (compile-assembly #<glil (program () (source ((breakpoin...>)
   194:  6 (glil->assembly #<glil (program () (source  
((breakpoi...> ...)
   186:  5 (lp (#<glil (program () (source ((breakpoint . #...> #)  
# ...)
   194:  4 (glil->assembly #<glil (program () (source  
((breakpoi...> ...)
   186:  3 (lp (#<glil (const #2f64((0.0 1.0) (2.0 3.0)))> # # #) # ...)
   377:  2 (glil->assembly #<glil (const #2f64((0.0 1.0) (2.0  
3....> ...)
   120:  1 (assoc-ref-or-acons ((#<<variable-cache-cell> k...> .  
1)) ...)
         0 (assoc-ref ((#<<variable-cache-cell> key: f!> . 1)) #)

But these three

((cut f! #2f64((0 1) (2 3)) <>) x)
(cut f! #f64(0 1 2) <>)
((cut f! #f64(0 1 2) <>) x)

all work. No error with ,option interp #t.

Regards,

	Daniel

Re: compiler bug in 1.9.9

[Andy Wingo]

On Tue 30 Mar 2010 17:02, Daniel Llorens del Río <daniel.llorens@bluewin.ch> writes:

>         0 (assoc-ref ((#<<variable-cache-cell> key: f!> . 1)) #2f64((0 1) (2 3)))

This was actually a problem with the runtime, not the compiler: equal?
on an array and a non-array bombed. Fixed, thanks!

Andy
-- 
http://wingolog.org/

Re: compiler bug in 1.9.9

[Daniel Llorens del Río]

On 30 Mar, 2010, at 21:24, Andy Wingo wrote:

> On Tue 30 Mar 2010 17:02, Daniel Llorens del Río  
> <daniel.llorens@bluewin.ch> writes:
>
>>         0 (assoc-ref ((#<<variable-cache-cell> key: f!> . 1)) #2f64 
>> ((0 1) (2 3)))
>
> This was actually a problem with the runtime, not the compiler: equal?
> on an array and a non-array bombed. Fixed, thanks!

Thanks for this, and also for the autocompile fixes. I've got another  
one, again with the latest git.

The first two lines work but the third one… doesn't:

 > (call-with-values (lambda () (values #s32() #f64(0))) list)
=> (#s32() #f64(0.0))

 > (call-with-values (lambda () (values #s32(0) #f64())) list)
=> (#s32(0) #f64())

 > (call-with-values (lambda () (values #s32() #f64())) list)
=> (#s32() #s32())

Although this one also works…

 > (call-with-values (lambda () (f #s32() #f64())) (lambda (a b)  
(format #f "~a ~a" a b)))
=> "#s32() #f64()"

Regards,
	
	Daniel

Re: bug in 1.9.9

[Daniel Llorens del Río]

On 31 Mar, 2010, at 17:55, Daniel Llorens del Río wrote:

> Although this one also works…
>
> > (call-with-values (lambda () (f #s32() #f64())) (lambda (a b)  
> (format #f "~a ~a" a b)))
> => "#s32() #f64()"

Sorry, this should be  … (lambda () (values #s32() #f64())) …

Regards,

	Daniel